Located between urban area and rural area, urban–rural fringe is challenged with urbanization related social-ecological problems. Accurately identifying the urban–rural fringe can help to integrated urban–rural development planning, especially in metropolitan region. Among the various case studies to identify the urban–rural fringe, land use degree and impervious surface area are widely used. However, both indexes are only focused on land development size, resulting in coarse identifying results.
It is aimed to propose a three-dimensional approach to integrating land development size, pattern and density, in order to accurately identifying the urban–rural fringe.
Landsat TM and DMSP/OLS datasets were used to establish a three-dimensional index system consisting of land development size (LDS), land development pattern (LDP) and land development density (LDD). Self-Organizing Feature Map (SOFM) was applied to identify the urban–rural fringe of Beijing City, China.
From 2001 to 2009, the inner boundary of the urban–rural fringe had expanded to outside the fifth ring road. Likewise, the outer boundary moved from the fifth to the sixth ring road. The new urban development zone was the main area of urban expansion controlled by urban planning, where the increments of urban–rural fringe was 1273.5 km2, accounting for 75.24% of the whole city. Partial correlation analysis indicated that LDS played a leading role in SOFM clustering, but the spatial continuity of the urban–rural fringe was the best when it was integrated with LDP and LDD, especially the latter to comprehensively define and quantify land development intensity.
The integration of land development size, pattern and density is effective to quantify land development intensity, and thus to identify the urban–rural fringe in metropolitan regions.